Processed GPM-DPR Convective Profiles (April 2014 - November 2023)
Data files
Jan 10, 2025 version files 97.55 GB
-
deep_convection_2014.nc
4.81 GB
-
deep_convection_2015.nc
6.40 GB
-
deep_convection_2016.nc
6.41 GB
-
deep_convection_2017.nc
6.36 GB
-
deep_convection_2018.nc
10.60 GB
-
deep_convection_2019.nc
13.08 GB
-
deep_convection_2020.nc
12.60 GB
-
deep_convection_2021.nc
12.74 GB
-
deep_convection_2022.nc
13.16 GB
-
deep_convection_2023.nc
11.39 GB
-
README.md
8.69 KB
Abstract
This is the processed dataset that corresponds with the paper titled: "A multi-frequency spaceborne radar perspective of deep convection" in AMS JAMC. The paper looks to evaluate the information content of shorter wavelengths of radar observed within deep convection collected by NASA's Global Precipitation Measurement mission (GPM) Dual-frequency Precipitation Radar (DPR). This dataset is the colocated and quality controlled radar profiles measured at Ku- (13 GHz) and Ka- (35 GHz) band within convection labeled by the GPM-DPR algorithms. These profiles occur globally from about -65 S and 65 N. The dataset is broken into yearly files, where each one contains radar profiles of shape n_samples, range_from_radar. Inside the files are also the original data (not-QCed), the temperature profile and a bunch of derived metrics from the profiles (e.g., echotops that were used in the paper). Overall, there are more than 39 million profiles in the dataset observed from March 2014 to November 2023. The script used to make these files is also included.
README: Processed GPM-DPR Convective Profiles (April 2014 - November 2023)
https://doi.org/10.5061/dryad.g4f4qrg0w
Description of the data and file structure
The data here are quality controlled and extracted GPM-DPR profiles from within convection globally. Each File contains one calendar year of data. In total there are more than 30 million profiles included.
Files and variables
Common between files:
Dimensions: this varies depending on the variable, but in general there is the nsample dimension which are individual profiles of data. range is the vertical dimension, which is sampled to 250 m vertical bins. freq is the different frequencies of the DPR, 0 is the Ku-band, 1 is the Ka-band.
Variables: Here is the list of all variables in the files.
- R: Near surface precipitation rate from the 2A.DPR file. mm/hr
- lon: longitude of the surface gate of the profile. degrees
- lat: latitude of the surface gate of the profile. degrees
- height: the height of the vertical range bin above mean sea level. meters
- echotop: the heightStormTop field from the 2A.DPR file. This is roughly the 12 dBZ echotop. meters
- surface_type: The surface type used in the 2B.CMB file (original location: hdf['KuKaGMI']['Input']['surfaceType'])
- surface_R_cmb: the near surface precipitation rate from the 2B.CMB retrieval. mm/hr
- surface_Rsigma_cmb: the uncertainty in the 2B.CMB retrieval. mm/hr
- z_raw: the measured reflectivity from the 2A.DPR file (zFactorMeasured) without any quality control (incudes the surface echo and the noise above echotop. dBZ
- z_meas: the measured reflectivity from the 2A.DPR file (zFactorMeasured) with the above echotop noise removed and the surface echo. dBZ
- z_corr: the retrieved attenuation corrected reflectivity from the 2B.CMB retrieval. dBZ
- z_retr: the retrieved Ka-band reflectivity WITHOUT attenuation correction. dBZ. This was calculated using code from Mircea Grecu NASA
- z_meas_ice: same as z_meas, but isolated to the ice layer using the melting layer height. dBZ
- z_corr_ice: same as z_corr, but isolated to the ice layer using the melting layer height. dBZ
- z_retr_ice: same as z_retr, but isolated to the ice layer using the melting layer height. dBZ
- z_meas_rain: same as z_meas, but isolated to the rain layer using the melting layer height. dBZ
- z_corr_rain: same as z_corr, but isolated to the rain layer using the melting layer height. dBZ
- z_retr_rain: same as z_retr, but isolated to the rain layer using the melting layer height. dBZ
- IWC: the ice water content from the 2B.CMB retrieval. g/m^3
- RWC: rain water content from the 2B.CMB retrieval. g/m^3
- Nw: the normalized intercept parameter from the 2B.CMB retrieval. log(m^4)
- Dm: the mass weighted mean diameter from the 2B.CMB retrieval. mm
- Nw_rain: the normalized intercept parameter from the 2B.CMB retrieval isolated to the rain layer. log(m^4)
- Dm_rain: mass weighted mean diameter from the 2B.CMB retrieval isolated to the rain layer. mm
- Nw_ice: the normalized intercept parameter from the 2B.CMB retrieval isolated to the ice layer. log(m^4)
- Dm_ice: mass weighted mean diameter from the 2B.CMB retrieval isolated to the ice layer. mm
- airTemperature: temperature at each gate from the 2A.DPR file (hdf['FS']['VER']['airTemperature']). K
- PIA: path integrated attenuation from the 2A.DPR file (hdf['FS']['SLV']['piaFinal']) dB
- PIA_SIM: path integrated attenuation from the 2B.CMB file for only the Ka-band. dB
- her_ka_z1: the Ku-band reflectivity at the maximum Ka-band reflectivity above the melting layer. dBZ
- her_ka_z2: the Ka-band reflectivity at the maximum Ka-band reflectivity above the melting layer. dBZ
- her_ka_h: the height of the maximum Ka-band reflectivity above the melting layer. meters above mean sea level
- dfr_et20_ku: the dual-frequency ratio (Ku-Ka) at the Ku-band 20 dBZ echotop height. dB
- dfr_et30_ku: the dual-frequency ratio (Ku-Ka) at the Ku-band 30 dBZ echotop height. dB
- scan_idx: the scan index. Can be used to isolate near nadir scans.
- dtime: the datetime of the sample
- MSindex: the multiple scattering flag from the 2A.DPR file.
- MSindexKu: the multiple scattering flag from the 2A.DPR file for Ku-band
- MSindexKa: the multiple scattering flag from the 2A.DPR file for the Ka-band
- MSkneeDFRindex: the multiple scattering flag from the 2A.DPR file for the DFR knee
- et12_ku: the Ku-band 12 dBZ echotop calculated by our code. meters above sea level
- et20_ka: the Ka-band 20 dBZ echotop calculated by our code. meters above sea level
- et20_ku: the Ku-band 20 dBZ echotop calculated by our code. meters above sea level
- et30_ka: the Ka-band 30 dBZ echotop calculated by our code. meters above sea level
- et30_ku: the Ku-band 30 dBZ echotop calculated by our code. meters above sea level.
- et40_ku: the Ku-band 40 dBZ echotop calculated by our code. meters above sea level.
- her_ku_z1: the Ku-band reflectivity at the maximum Ku-band reflectivity above the melting layer. dBZ
- her_ku_z2: the Ka-band reflectivity at the maximum Ku-band reflectivity above the melting layer. dBZ
- her_ku_h: the height of the maximum Ku-band reflectivity above the melting layer. meters above mean sea level
- melting_height: the melting layer height from the 2B.CMB file. meters above sea level
- dfr_et40_ku: the dual-frequency ratio at the 40 dBZ Ku-band echotop height. dB
- dfr_et20_ka: the dual-frequency ratio at the 20 dBZ Ka-band echotop height. dB
- dfr_et30_ka: the dual-frequency ratio at the 30 dBZ ka-band echotop height. dB
- clutter_height: the minimum clutter free echo height from the 2A.DPR file. meters above sea level.
- surface_elevation: the elevation of the surface from the 2A.DPR file. meters above sea level.
- t_et12_ku: temperature at the 12 dBZ Ku-band echotop height. K
- t_et20_ka: temperature at the 20 dBZ Ka-band echotop height. K
- t_et20_ku: temperature at the 20 dBZ Ku-band echotop height. K
- t_et30_ka: temperature at the 30 dBZ Ka-band echotop height. K
- t_et30_ku: temperature at the 30 dBZ Ku-band echotop height. K
- t_et40_ku: temperature at the 40 dBZ Ku-band echotop height. K
File: deep_convection_2014.nc
Description: All convective profiles from April 2014 through December 2014 collected only in the inner part of the GPM-DPR swath, where both Ku- and Ka-band data are available.
File: deep_convection_2015.nc
Description: All convective profiles from January 2015 through December 2015 collected only in the inner part of the GPM-DPR swath, where both Ku- and Ka-band data are available.
File: deep_convection_2016.nc
Description: All convective profiles from January 2016 through December 2016 collected only in the inner part of the GPM-DPR swath, where both Ku- and Ka-band data are available.
File: deep_convection_2017.nc
Description: All convective profiles from January 2017 through December 2017 collected only in the inner part of the GPM-DPR swath, where both Ku- and Ka-band data are available.
File: deep_convection_2018.nc
Description: All convective profiles from January 2018 through May 2018 collected only in the inner part of the GPM-DPR swath, where both Ku- and Ka-band data are available. Then May 2018 to December 2018 is the full swath because of the scan pattern change in May 2018 to the Ka-band.
File: deep_convection_2019.nc
Description: All convective profiles from January 2019 through December 2019.
File: deep_convection_2020.nc
Description: All convective profiles from January 2020 through December 2020.
File: deep_convection_2021.nc
Description: All convective profiles from January 2021 through December 2021.
File: deep_convection_2022.nc
Description: All convective profiles from January 2022 through December 2022.
File: deep_convection_2023.nc
Description: All convective profiles from January 2023 through November 2023. This ends at the orbit boost in November 2023.
Code/software
You should be able to use any software than can load netCDF4 files. These were made using python and xarray. Example way to load a file:
import xarray as xr
ds = xr.open_dataset('deep_convection_2023.nc')
Access information
Data was derived from the following sources:
- 2A.DPR: 10.5067/GPM/DPR/GPM/2A/07
- 2B.CMB: 10.5067/GPM/DPRGMI/CMB/2B/07
Methods
These files are a combination of the 2A.DPR files and the 2B.CMB files from GPM-DPR (see this webpage for the raw files: https://gpm.nasa.gov/data/directory). They are acquired to enable the use of both the uncorrected measured radar reflectivities (2A.DPR) and the combined radar-radiometer retrievals of precipitation rate and water content (Grecu et al. 2016; Olson 2022). Data are sub-sampled to only convectively labeled profiles from the GPM algorithm (i.e., Awaka et al. 2021) and profiles that extend at least 5 km above the surface elevation, where the top was determined from the GPM storm top algorithm (heightStormTop). These criteria were used to isolate deep convection. Note that both the raw measured profiles as well as the quality controlled profiles (i.e., surface echo and noise above cloud top removed) are included in the dataset.